Junk 30, 1899.] 



SCIENCE. 



895 



present status. But this neglects an essen- 

 tial factor. The question really hinges on 

 tlie proportion of potential energy convertible 

 into heat which remained within the earth 

 when full grown. There is no great differ- 

 ence between the alternative hypotheses so 

 far as the amount of sensible heat at the 

 beginning of the habitable stage is con- 

 cerned. For, on the one hand, the white- 

 hot earth must have become relatively cool 

 on the exterior before life could begin, and, 

 on the other, it is necessary to assume a 

 sufiBciency of internal heat coming from 

 impact and intei-nal compression, or other 

 changes, to produce the igneous and crys- 

 talline phenomena which the lowest rocks 

 present. The superficial and sub-super- 

 ficial temperatures in the two cases could 

 not, therefore, have been widely different. 



So far as the temperatures of the deep 

 interior are concerned there is only recourse 

 to hypothesis. It is probable that there 

 would be a notable rise of temperature 

 toward the center of the earth in either 

 case. In a persistently liquid earth this 

 high central temperature would be lost 

 through convection, but if central crystalli- 

 zation took place at an early stage through 

 pressure, much of the high central heat 

 might be retained. In a meteor-built earth, 

 solid from the beginning, very much less 

 convectional loss would be suffered, and the 

 central temperature would probably corre- 

 spond somewhat closely to the density. The 

 probabilities, therefore, seem somewhat to 

 favor a higher thermal gradient toward the 

 center in the case of the solid meteor- built 

 earth. 



But if we turn to the consideration of 

 potential energy, there is a notable differ- 

 ence between the two hypothetical earths. 

 In the liquid earth, the material must be 

 presumed to have arranged itself according 

 to its specific gravity, and, therefore, to 

 have adopted a nearly complete adjustment 

 to gravitative demands ; in other words, to 



have exhausted, as nearly as possible, its po- 

 tential energy, i. e., its ' energy of position.' 

 On the other hand, in an earth built up 

 by the accretion of meteorites without free 

 readjustment there must have been initially 

 a heterogeneous arrangement of the heavier 

 and lighter material throughout the whole 

 body of the earth, except only so far as the 

 partial liquefaction and the very slow, 

 plastic, viscous and diifusive rearrangement 

 of the material permitted an incipient ad- 

 justment to gravitative demands. A large 

 amount of potential energy was, there- 

 fore, resti-ained, for the time being, from 

 passing into sensible thermal energy. This 

 potential energy thus restrained is sup- 

 posed to have gradually become converted 

 into heat as local liquefaction and viscous, 

 molecular and massive movements per- 

 mitted the sinking of the heavier material 

 and the rise of the lighter material. This 

 slow conversion of potential energy into 

 sensible heat is thought to give to the slow- 

 accretion earth a very distinct superiority 

 over the hot liquid earth when the com- 

 bined sum of sensible and potential heat is 

 considered. The theoretical difference is 

 capable of approximate computation, and 

 Mr. F. R. Moulton has kindly undertaken 

 to make the computation in a simplified 

 hypothetical case which may give some im- 

 pression of the possible order of magnitude 

 of this factor. For the purposes of the 

 computation the earth was assumed to have 

 been composed of 40 % of metal with a 

 normal surface specific gravity of 7 and 

 60 % of rock with a normal surface specific 

 gravity of 3. These combined would give 

 an earth whose average specific gravity 

 would be only 4.6. The real specific grav- 

 ity (5.6) is supposed to have been obtained 

 by compression which would amount to 

 about 18 %. Very likely the proportion 

 of metal is put too high and the effect of 

 compression too low, but the purpose of the 

 computation is only to show the theoretical 



